optimization on pricing and overconfidence problem in a
TRANSCRIPT
University of Southern Denmark
Optimization on pricing and overconfidence problem in a duopolistic supply chain
Xu, Lei; Shi, Xiaoran; Du, Peng; Govindan, Kannan; Zhang, Zhenchao
Published in:Computers and Operations Research
DOI:10.1016/j.cor.2018.04.003
Publication date:2019
Document version:Accepted manuscript
Citation for pulished version (APA):Xu, L., Shi, X., Du, P., Govindan, K., & Zhang, Z. (2019). Optimization on pricing and overconfidence problem ina duopolistic supply chain. Computers and Operations Research, 101, 162-172.https://doi.org/10.1016/j.cor.2018.04.003
Go to publication entry in University of Southern Denmark's Research Portal
Terms of useThis work is brought to you by the University of Southern Denmark.Unless otherwise specified it has been shared according to the terms for self-archiving.If no other license is stated, these terms apply:
• You may download this work for personal use only. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying this open access versionIf you believe that this document breaches copyright please contact us providing details and we will investigate your claim.Please direct all enquiries to [email protected]
Download date: 22. Apr. 2022
Accepted Manuscript
Optimization on pricing and overconfidence problem in a duopolisticsupply chain
Lei Xu , Xiaoran Shi , Peng Du , Kannan Govindan ,Zhenchao Zhang
PII: S0305-0548(18)30086-8DOI: 10.1016/j.cor.2018.04.003Reference: CAOR 4444
To appear in: Computers and Operations Research
Received date: 24 November 2015Revised date: 30 December 2017Accepted date: 4 April 2018
Please cite this article as: Lei Xu , Xiaoran Shi , Peng Du , Kannan Govindan , Zhenchao Zhang ,Optimization on pricing and overconfidence problem in a duopolistic supply chain, Computers andOperations Research (2018), doi: 10.1016/j.cor.2018.04.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a serviceto our customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, andall legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
1
Highlights:
Put forward the retailer’s overconfidence in a duopolistic supply chain
Categorize two cases on retailer’s overconfidence information and derive the retailer’s optimal pricing and
ordering decision in Nash game
Extension to a Stackelberg-like model to investigate the impact of first-mover advantage on the outcome of
overconfidence
Overconfidence does not necessarily damage the supply chain performance
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
2
Optimization on pricing and overconfidence problem in a duopolistic
supply chain
Lei Xu a,b
; Xiaoran Shi a; Peng Du
a*; Kannan Govindan
c; Zhenchao Zhang
a
a Research Center for Circular Economy and Sustainable Development, School of Management, Tianjin University of
Technology, Tianjin 300384, China; [email protected]; [email protected] b Business School, Nankai University, Tianjin 300071, China c Center for Engineering Operations Management, Department of Technology and Innovation, University of Southern
Denmark, Denmark DK-5230 Odense M, Denmark; [email protected] * Correspondence: [email protected]
Abstract We analyze the impact of the retailer’s overconfidence on the supply chain performance. We
consider a duopolistic market with uncertain demand where one overconfident retailer and one rational retailer
compete in selling the same product. We identify two cases in terms of the rational retailer’s overconfidence
awareness: (1) the rational retailer is aware of the other retailer’s overconfidence; (2) the rational retailer is not
aware of the other retailer’s overconfidence. We first analyze the two retailers’ decision processes in a Nash
game and derive their optimal decisions, and further discuss a Stackelberg-like game for an extension. We
conduct a numerical analysis to examine the impact of overconfidence on the optimal pricing, ordering
decisions and expected profits. We find that the higher the overconfident level is, the higher selling price that
the overconfident retailer charges. Together, the results stress that the overconfidence does not necessarily
damage the supply chain performance.
Keywords: Supply chain management, Overconfidence, Duopolistic market, Overconfidence awareness
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
3
1. Introduction
As one of the most commonly seen irrational behaviors in decision making, overconfidence has drawn much
attention and been widely acknowledged having significant impacts in the economic and management areas.
Nowadays, it is ubiquitous that retailing giants cut prices down to compete with their rivals in an aggressive
way, even behave irrationally and fight like Kilkenny cats, due to their overconfidence on market outlooks.
Although there is no general tendency toward overconfidence (Fowler and Johnson, 2011), previous market
share or profit level, even their finance capital etc. may become anchors or reference points in their decision
making, particularly their prominent successes in such business performances would bloat them to behave
irrationally either trigger a price war in their markets or break into new markets. For example, with previous
rapid expansion and ambition on booming market share, JD.com has successively launched two price wars in
three years, with Dangdang.com in 20101 and with Gome.com.cn and Suning.com in 2012
2, yet benefitted
from neither as wished. In contrast, the price war launched optimistically by Gree Electric Appliances in 2014
helped expand the market share and digest finished products inventory dramatically. It also happens in other
markets, a well-known case is that in the crowfunding market, MI Tech. Ltd. and PepsiCo Corp. almost
simultaneously initiated their next generation smartphone in crowdfunding platform, the latter yet finally
failed with an ambitious funding target 3. In these cases above, the firms behave in overconfidence to some
degree in their decision makings. However, the outcomes are not always the same.
Plous (1993) notes that “no problem in judgment and decision making is more prevalent and
more potentially catastrophic than overconfidence”.
From a traditional perspective, overconfidence is perceived as a sabotaging factor in a wide range of
areas: unprofitable merger and acquisition (Malmendier & Tate, 2005), mal-investment in secondary market
(Heaton, 2002), even unwise channel entry (Camerer & Lovallo, 1999) and inventory imbalance (Ren &
Croson, 2013) in operational level. Our study challenges the conventional wisdom that overconfidence makes
decision maker (even supply chain) worse off by examining the impact of overconfidence in the context of
duopolistic market. Specifically, our objective is to assess the impact of one retailer’s overconfidence in a
duopolistic supply chain and how the rival’s overconfidence awareness adjusts this impact in different market
power conditions. Our research question is motivated by the observation that the outcomes of overconfidence
in duopolistic market have been confirmed to vary substantially in industries.
To address the research question, we employ a series of models with different conditions of the rival’s
overconfidence awareness and compare the outcomes. The models consist of a supply chain with one supplier
and two competing retailers, one retailer holds overconfident biased belief on market information, and the
1 For details, refer to http://tech.ifeng.com/internet/special/360buypkdangdang/ 2 Refer to http://tech.hexun.com/2012/jiagezhan2012/ 3 PepsiCo Corp. initiated a project for Pepsi Phone in Nov 2015, and ambitiously set his funding target as 3 million RMB with an
aggressively low price at 499 RMB. Ultimately this project failed with a poor fund-raising. Almost at the same time, MI Tech. Ltd.
also initiated a project for its new generation smartphone with a lower funding target of 1 million and a relatively higher unit selling
price of 1000 RMB, the project finally raised more than 35.595 times of the initial funding target. Refer to
http://z.jd.com/project/details/30677.html
and https://izhongchou.taobao.com/dreamdetail.htm?spm=a215p.128754.653087.26.4dfb51b6JYu Q2u&id= 10050396
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
4
other retailer is rational. Further, the models also hinge on the market power to test the robustness of the
results by assuming that the two retailers play Nash game and Stackelberg-like game respectively.
The paper proceeds as follows. We review the literature in §2. Following that, we describe the problem
and present our models in §3. Section 4 discusses two retailers’ pricing and ordering decisions in a duopolistic
market. Section 5 presents numerical studies to discuss the impact of overconfidence on pricing, ordering and
profits. Finally, Section 6 concludes and provides direction for future research.
2. Literature review
There are two primary streams of research that relate to our analysis, which are studies conducted on pricing
competition among retailers and studies on overconfidence, respectively. For the first stream, the competition
among retailers is extensively addressed in academia and analyzed from different perspectives. Ingene and
Parry (1995) analyze how two competing retailers make their pricing strategies when products are
homogenous. Bernstein and Federgruen (2003) extend to address two competing retailers’ pricing strategies in
Bertrand and Cournot game. Xiao and Qi (2008) consider a scenario of pricing competition between two
retailers where the manufacturer’s production cost is disrupted. Parthasarathi et al. (2011) study how two
competing retailers make pricing and ordering decisions when the manufacturer acts as a Stackelberg leader.
More recently, scholars begin to analyze the dual channel problem. Mukhopadhyay et al. (2011) consider two
firms sell complementary goods in a Stackelberg game under information asymmetry. Chen et al. (2013)
consider the impact of competing for price on the supply chain in Nash game and Stackelberg game,
respectively. Du et al. (2016) discuss the pricing competition between the incumbent and entrant firm on the
innovative product. A literature review on the pricing competition and its extensions can be found in Agatz et
al. (2008). Although many of these papers consider the pricing competition of the firms, none addresses the
retailer’s decision preferences in a competing retailing environment. As a complement to the literature
reviewed above, our paper explores the impact of overconfidence on pricing decision in a duopolistic supply
chain context.
Decision-makers’ behavior preference, the second stream, also gains great popularity in academia.
Pertinent scholars have targeted their concerns on risk aversion (Yang et al., 2009), loss-averse (Liu et al.,
2013), fairness concern (Cui et al., 2007; Ho et al., 2014), anchoring effects (Chandrashekaran and Grewal,
2006), the opportunism behavior (Hawkins et al., 2008), and spin off effect (Klepper and Sleeper, 2005),
amongst many others. However, to the best of our knowledge, there has been limited research on the
overconfidence in OM and MS area, whereas our paper considers the overconfidence in a supply chain with
two retailers. The new trend is that scholars begin to pay close attention to overconfidence in its various forms
to examine its impact and describe a variety of behavior in a supply chain environment (Croson et al., 2011;
Ren and Croson, 2013).
Research on overconfidence is quite mature in the field of psychology and has been widely explored in
the field of behavioral finance and behavior management. Particularly, Gervais et al. (2002) find that
overconfident managers can make better decisions than rational managers in coordinating the interests of the
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
5
shareholders. Heaton (2002) thinks that managers’ overconfidence leads to the low efficiency of investment
decision and then damages the interests of the enterprise. Nowadays, there is some literature starting to
investigate the overconfidence in OM/MS area. Croson et al. (2011) summarize previous experimental study
on human preference, and use salvage costs and price adjustments to revise the overconfident retailer’s
ordering decisions. Ren and Croson (2013) show that the order bias could be caused by the overconfidence
using a preference test. In a similar vein, Proeger and Meub (2014) show that overconfidence is a social rather
than an individual bias through an experiment conducted in their study. Moreover, Wang et al. (2015) build a
model of a retailer who is overconfident on mean and variance of the demand and discuss the effectiveness of
supply contract in the supply chain coordination. Ren et al. (2017) conduct a theoretical research as an
extension of their study (Ren and Croson, 2013). They model the newsvendor’s overprecision behavior as a
biased belief that the distribution of demand has an underestimated variance. Nevertheless, these papers only
consider a primary market containing one retailer, or a supply chain with one manufacturer and one retailer.
Furthermore, experiment methods are always used in these studies to test and verify the overconfidence in the
supply chain. Differing from most of the existing literature, under random and price-dependent demand, our
paper analytically investigates different pricing games between one rational retailer and one overconfident
retailer, and addresses the impact of overconfidence on supply chain performance under different
overconfident information settings. More recently, Li et al. (2016) study the ordering game between two
overconfident newsvendors. In comparison, our paper not only explores the impact of one retailer’s
overconfidence on both retailers’ pricing and ordering decisions, but also examines how the rational retailer’s
knowledge about his rival’s overconfidence adjusts this impact.
3. Problem descriptions
Consider a duopolistic supply chain wherein a rational decision-maker retailer 1 and retailer 2 with
overconfident bias on the market demand compete in Nash game on selling prices under uncertain demand.
According to retailer 1’s awareness about retailer 2’s overconfidence, we study two cases in the following: (1)
retailer 1 is aware of retailer 2’s overconfidence; and (2) retailer 1 considers retailer 2 as a rational decision-
maker as well. The manufacturer produces products with unit cost , the two retailers place their orders
( ) with a unit price and then sell products to consumers with a price , respectively. The salvage
value for the unsold products is per unit. The parameters used above have the following relationship:
.
In the duopolistic market, the demand is stochastic and price-dependent. Similar to the study of Mills
(1959), we assume that the demand function is ( ) ( ) . Therefore, the retailers’ actual
demand functions are
( ) , (1)
( ) , (2)
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
6
where is the market scale of retailer , is the price sensitivity and is the cross-price sensitivity
satisfying . Moreover, is a continuous random variable on the range of with mean and
variance , and has a cumulative distribution function ( ) and probability density function ( )( ).
For the tractability, it is assumed that follows a uniform distribution. Besides, to avoid trivial cases, it is
assumed (Petruzzi and Dada, 1999).
Referring to the classification of overconfidence defined by Moore and Healy (2008), the retailer’s
overconfident bias belongs to overprecision. According to Ren and Croson (2013), we adopt a mean-
preserving but variance-reducing transformation of the actual market demand. Therefore, in the view of
retailer 2, the demand functions he faces and retailer 1 faces are shown in (3) and (4), respectively,
( ) ( ) , (3)
( ) ( ) , (4)
where indicates the overconfident level and satisfies . Here, the smaller is, the higher the
overconfident level is. Especially, when , retailer 2 has no overconfidence on demand variance (i.e., he
is a rational decision-maker). Meanwhile, in the view of retailer 2, the mean of the market demand is
( ) ( ) and the variance is ( ) ( ).
Based on the demand functions above, we aim to analyze two competing retailers’ pricing and ordering
decisions in two cases. To clarify our exposition, (i) we use expected profit to represent the actual expected
profit of manufacturer and retailers, and anticipated profit to represent the fictitious expected profit of
manufacturer and retailers, respectively; (ii) we use the subscripts 1 and 2 to represent retailer 1 and 2, and use
superscripts , , , and to represent the benchmark case, Case , Case , Case , and Case ,
respectively throughout the paper. To be specific, we first consider a Nash game, in which retailer 1 realizes
retailer 2’s overconfidence in Case I while retailer 1 is not aware of retailer 2’s overconfidence in Case II.
Then, as an extension, a Stackelberg game (retailer 2 is the leader) is studied in Case III and IV. Similar to
Case I and Case II, retailer 1 has a foresight on retailer 2’s overconfident behavior in Case III, while he
considers retailer 2 as a rational decision maker in Case IV. In order to provide a consistent notation
throughout this paper, we briefly describe the parameters and variables appeared in the derivation of the
theorem and corollaries, as shown in Table 1.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
7
Table 1 List of Symbols
production cost per unit wholesale price per unit
salvage value for unsold product per unit selling price per unit
market scale of retailer the price sensitivity
cross-price sensitivity a continuous random variable in the support of
( ) cumulative distribution function of ( ) probability density function of
mean of variance of
retailer ’s profit when the two retailers are
rational
the optimal pricing of retailer in case K
( ) retailer ’s expected profit in case the profit of the entire supply chain in case K
represents the optimal pricing of retailer in case K in retailer ’s view
represents the optimal ordering quantity of retailer in case K in retailer ’s view
represents the profit of retailer in case K in retailer ’s view
( ) represents the demand of retailer in retailer ’s view which is related with .
4. The analysis of two cases
It is known that the retailers’ objectives are to maximize their own profits by choosing prices and ordering
quantities. As a benchmark where retailer 1 and 2 are both rational, retailer i’s expected profit is ( ).
4.1 Case I
In this case, both retailers make their pricing and ordering decisions simultaneously. Retailer 1 knows that
retailer 2 is overconfident on his demand variance and the overconfidence level; however, in retailer 2’s view,
they both are rational, thus their market demand functions are equations (3) and (4). That is, retailer 2 does not
recognize his own overconfident behavior. Accordingly, we build both retailers’ profit functions in retailer 2’s
view as follows.
Retailer 1’s profit function is
(
) {
(
)
(
) (
)
(
)
.
Retailer 2’s profit function is
(
) {
(
)
(
) (
)
(
)
.
Since each retailer seeks to maximize his own profit, in the view of retailer 2, the optimal prices are
shown in Theorem 1.
Theorem 1. In the view of retailer 2, the optimal selling prices are
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
8
( )( ) (
)
( ) ,
( )( ) (
)
( ) ,
where
( ) ( )
and
( ) ( )
.
Given the values of ,
, , and
, the optimal ordering, from retailer 2’s perspective, are
( )
and
( )
,
respectively.
Retailer 2 believes he can predict retailer 1’s optimal pricing and ordering decisions accurately, and
maximize his own profit accordingly. However, he is overconfident on the market demand variance and
makes a distorted pricing decision, which would lead to market disruption, thereby causing a decrease in both
his profit and profits of retailer 1.
Compared with retailer 2, retailer 1 can predict the market demand accurately. Meanwhile, he knows
retailer 2’s overconfident level and decision process clearly. Thus, retailer 1 can predict retailer 2’s optimal
pricing and ordering decisions accurately, which embodies the value of information (Yu et al., 2001).
Retailer 1’s profit function is given as follows,
(
) {
(
)
(
) (
)
(
)
(5)
Based on the analysis above, the best response of retailer 1 is shown in Theorem 2.
Theorem 2. Retailer 1’s optimal selling price is
( )( ) (
)
( ) ,
where
( ) ( )
and is retailer 2’s optimal price predicted by retailer 1.
Besides, retailer 1’s optimal ordering quantity is
, where is retailer
2’s optimal pricing predicted by retailer 1. Here, note that
in case I, which means retailer 1 can
predict retailer 2’s pricing decision accurately.
4.2 Case II
Different from case I, in case II, retailer 1 is not aware of retailer 2’s overconfidence. In other words, retailer 1
thinks that retailer 2 has the same knowledge of demand as he does.
4.2.1 The rational retailer problem
As illustrated at the beginning of this section, retailer 1 thinks that both of them are rational and their demand
functions have been given in (1) and (2). We utilize Theorem 3 to demonstrate the optimal pricing strategy in
retailer 1’s point of view.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
9
Theorem 3. In the view of retailer 1, the optimal pricings are
( )( ) (
)
( ),
( )( ) (
)
( ),
where
( ) ( )
and
( ) ( )
.
In the view of retailer 1, the optimal ordering quantities of both retailers are
,
, respectively, where the values of ,
, and
are
shown in Theorem3.
4.2.2 The overconfident retailer problem
As mentioned above, retailer 2 thinks that both of them are rational and the market demands are given in
equations (3) and (4), respectively. He predicts his own optimal pricing and ordering decisions. Meanwhile, he
believes that his rival has the same prediction and also considers the market demand functions to follow
equations (3) and (4). Similar to the results obtained in section 4.2.1, retailer 2 makes the same decisions.
Corollary 1. The optimal pricing is decreasing with respect to .
The result of Corollary 1 implies that the higher retailer 2’s overconfident level is, in retailer 2’s view,
the smaller the demand fluctuation is and the more precision his prediction is.
Corollary 2. is strictly convex with respect to
.
This corollary reveals the relationship between the prices of retailers 1 and 2 in case I, it tells us that in
retailer 1’s view, he could adjust his rival’s selling price by designing his selling price, and because of the
convexity, the results implies that retailer 1 has two pricing options for the same price of retailer 2. However,
he can choose the lower one to capture the competitive pricing advantage.
Corollary 3. Retailer 2’s expected profit has the following properties,
1. ( ) ( );
2. ( ) ( ) increases with respect to the overconfident level.
Corollary 3 shows the negative effect of overconfidence which can lead to order bias and a loss of profit.
After that, we aim to analyze the relationship among ( ), (
), and ( ), which is given in
Corollary 4.
Corollary 4. ( ) (
) ( ).
Corollary 4 implies that retailer 1 can make more profit when competing with an overconfident retailer
than with a rational retailer. When retailer 1 realizes retailer 2’s overconfidence, he can benefit from the
information on retailer 2’s overconfidence.
4.3 Extension
4.3.1 A Stackelberg game problem
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
10
Our analysis has assumed that two competing retailers sell products in Nash game in a duopolistic market and
decide selling prices simultaneously. However, in reality, different retailers may have different market power
and the game leader always has the first-mover advantage. Given this knowledge, in this section, we consider
the case that the retailer who has greater market power is more prone to behave in overconfidence in the
duopolistic market environment. Hence both retailers interact in a Stackelberg-like game, where the
overconfident retailer (retailer 2) is the Stackelberg leader and the rational retailer (retailer 1) is the follower.
The sequence of events is defined as follows. Firstly, retailer 2 announces and to maximize his profit
(i.e., ). Then, in response to retailer 2’s pricing and ordering strategy, retailer 1chooses and to achieve
his profit (i.e., ) maximization.
Firstly, we discuss case III, in which retailer 1 knows that retailer 2 is overconfident. In the view of
retailer 2, retailer 1’s expected profit function is
(
) ∫ (
( ) )
( ) ∫ (
) ( ) *
( )
+
∫ (
( )
)
( ) .
Solving the first-order condition of retailer 1’s anticipated profit function with respect to and
,
respectively, we can obtain
( )( ) (
)
( ), (6)
( ) ( )
(7)
Due to the complexity of mathematical calculation, closed-form expressions for the optimal prices are
difficult to obtain and can’t be represented explicitly by . Therefore, by defining
( ) and
substituting equations (5) and (6) into retailer 2’s anticipated profit function, we can obtain
(
) ∫ (
( ) ) ( )
∫ ( ) ( )
( )
∫ (
( )
)
( )
∫ ( (
) ( ) ( )) ( )
∫ ( (
) ( )
)
( )
* ( )
( )
+ (8)
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
11
Taking the first-order condition of retailer 2’s anticipated profit function with respect to and ,
respectively, we get and
. Therefore, retailer 2’s optimal ordering quantity can be represented as
(
) ( )
.
As mentioned above, retailer 1 has predicted retailer 2’s decisions correctly before retailer 2 makes
decisions. Retailer 1’s optimal response (i.e., (
)) is obtained by setting
to find
and
. Finally, we can obtain that
( )(
) ( )
( ) and
( ) ( )
, and retailer 1’s optimal ordering quantity is
.
Then we discuss case IV: retailer 1 thinks retailer 2 is a rational decision-maker. Here, retailer 2’s
decision process is the same with case III, that is,
,
, and
. In the view of
retailer 1, retailer 2’s predictions on his decision variables ( and
) are
( )(
) (
)
( ) , (9)
(
) ( )
. (10)
When , the equations (10) and (11) are equal to the equations (13) and (14), respectively.
Meanwhile, retailer 2’s profit-maximizing function is a special case of equation (12). In a similar vein, let
and substitute equations (13) and (14) into retailer 2’s anticipated profit function ,
we can obtain
(
) ∫ (
(
)) ( )
∫ (
)
( )
Solving the first-order condition of retailer 2’s expected function with respect to and , respectively,
we can obtain
and
. Similarly, by setting
,
and
is derived, thereby retailer 1’s
maximum-profit response function
(
) can be obtained. In the next step, retailer 1’s optimal
pricing and ordering quantity are calculated as
( )(
) (
)
( ),
(
) ( )
and
, respectively. From the above results, retailer
2’s decisions keep the same in all four cases.
Previous literature suggests that the Stackelberg leader has the first-mover advantage, which is either on
the assumption of information symmetry or rational decision-makers. However, in this study, our results show
that retailer 1 has a late-mover advantage in case III. According to our analysis, when the leader is an
irrational decision-maker and the follower has the information that the leader is irrational, the leader never has
the first-mover advantage again, since the follower can make use of this information and benefit from it.
1 2( )p h p 12 2 1 2[ ( , , )]IVE z p p
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
12
4.3.2 Overconfidence on other market variables
So far we assume that the retailer is only overconfident in his estimation of the demand variance. In reality,
the retailer may be overconfident in the market scale and demand sensitivity. For example, the retailer may
overvalue the potential demand scale he faces. This kind of overconfidence is considered as overestimation,
which is different from overprecision (see the Introduction section). From this perspective, if retailer 2
overestimates the market scale, in his view, the demand function he faces is
( ) ( )
where ( ) indicates the overconfident level, and the larger is, the higher the overconfident level
is. Especially, when , retailer 2 has no overconfidence on the market scale.
The overconfident retailer may not only overestimate the market demand scale, but may also overrate
the demand variance in the overprecision type. That is, he has two kinds of overconfidence – overestimation
and overprecision. Under this circumstance, the market demand he faces can be expressed as
( ) ( ) ,
where and denote the level of the overestimation and overprecision, respectively. In particular, a larger
and a smaller indicate a higher overconfident level. Consequently, the mean of the market demand in
this case would be higher than that considered in the previous cases (i.e.,
( ) ( ) ( ) ), and the variance becomes smaller (i.e.,
( ) ( ) ). Since retailer 2’s demand expectation will increase, the
ordering quantity also increases, leading to an inventory problem. Meanwhile, as the two variables ( )
change simultaneously, both retailers’ decisions will be more complex.
Furthermore, the overconfident retailer overestimates his prediction on the price sensitivity. Therefore,
in the view of retailer 2, the demand function he faces is
( ) .
where ( )indicates the overconfident level, and the smaller is, the higher the overconfident level
is. Especially, when , retailer 2 has no overconfidence on the price sensitivity. In the view of retailer 2,
the demand increases (decreases) when he reduces (raises) the selling price.
The overconfident retailer’s optimal price is
( )( ) (
)
( ) , where
( ) ( )
.
Last but not least, the overconfident retailer may overestimate the demand scale and demand sensitivity,
and predict the demand variance in overprecision preference at the same time. Therefore, in the view of
retailer 2, the demand function he faces is
( ) ( ) ,
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
13
where , and indicate the overconfident level in terms of the demand scale, demand sensitivity and
demand variance, respectively. The smaller and is and the larger is, the higher the overconfident level
is. This overconfident model integrates all types of the retailer’s overconfidence over the demand completely.
For retailer 2, the overconfidence may promote him to make decision with a higher price and lower order
quantity, thereby leading to a smaller profit. Moreover, the cases of retailer 1’s knowledge on retailer 2’s
overconfidence become more complex.
Since the profit functions contain a large number of unknown parameters, it is difficult to compare the
optimal pricing, the optimal order quantity and maximal profit under different cases. Therefore, numerical
methods are utilized to conduct further analysis.
5. Numerical analysis
In this section, our study releases the assumption made in Section 3 to examine the robustness of the obtained
results and the impact of overconfident level on the supply chain performance. Here, detailed parameter
settings are referred to the existing studies (see Choi, 1991; Lau and Lau, 2003).
To capture the effects of the overconfidence and avoid the complex impacts of cost parameters, we select
the “actual” market demand function as
( ) ,
( ) .
where ( ) Other parameters are set as , , , respectively.
Considering the overconfidence, the demand functions in the view of retailer 2 turn to be
( ) , ( ) .
Given all these settings, the optimal price, ordering quantity and the maximal profit are solved via
Maple. When both retailers are rational, as defined in the benchmark model, the optimal prices are
and
, the optimal ordering quantities are
and
, the
maximum profits are and
, and the manufacturer’s maximum profit is
. In case II,
and .
5.1 The impact of overconfident level on optimal decisions
Fig.1 shows the relationships between the overconfident level and the optimal selling prices. With the increase
of retailer 2’s overconfident level, retailer 1 raises his selling price in case I. However, since retailer 1
considers retailer 2 as a rational decision-maker, his pricing strategy is not influenced by retailer 2’s
overconfidence in case II. In contrast, retailer 2 raises his selling price both in cases I and II. Results obtained
from the numerical analysis can also be verified in Corollary 1. Moreover, it is noticed that retailer 1 always
sets a lower selling price compared with retailer 2 (i.e.,
) to attract more customers in case I. From Fig.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
14
2, it is found that retailer 2’s optimal ordering quantity increases as increases in both cases I and II. On the
contrary, retailer 1’s optimal ordering quantity decreases with in case I and remains the same in case II.
Fig. 1. The effect of overconfidence on
optimal prices
Fig. 2. The effect of overconfidence on
optimal ordering quantities
Figs. 3 and 4 summarize the impact of the overconfidence on the profits of the supply chain. According
to the figures, retailer 1’s expected profit increases with the overconfident level. When it comes to retailer 2,
as he mistakenly predicts his rival’s decision, his expected profit decreases compared with the benchmark
model (i.e.,
,
). Furthermore, a higher overconfident level results in a higher selling price
and a lower ordering quantity for the overconfident retailer. Retailer 2’s anticipated profit is greater than the
actual expected profit, which is also verified in Corollary 2.
Fig. 3. The effect of overconfidence on profits
in case I
Fig. 4. The effect of overconfidence on profits
in case II
5.2 The value of information on overconfidence
For retailer 1, it is evident that his pricing and ordering strategies are different when he has or doesn’t have the
knowledge of retailer 2’s overconfidence. In order to reflect the value of this knowledge, we seek to analyze
both retailers’ profits in cases I and II, as shown in Tab. 2 and Fig. 5.
Table 2 The effect of overconfidence on retailer1’s profit in cases I and II
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
972.6176 970.6021 968.5783 966.5360 964.4856 962.4218 960.3430 958.2599 956.1566 954.0454 951.9160
972.5091 970.5136 968.5078 966.4817 964.4454 962.3937 960.3319 958.2496 956.152 954.0442 951.9160
0 0.2 0.4 0.6 0.8 1
1200
1600
2000
2400
a
prof
it I
1
I
2
I
22
I
M
0 0.2 0.4 0.6 0.8 1900
1300
1700
2100
2400
a
prof
it
II
1
II
2
II
M
a
I
1
II
1
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
15
Fig. 5. The effect of overconfidence on retailer 2’s profit
Tab. 2 shows that compared with case II, retailer 1 can obtain more profit in case I (i.e.,
). The
reason is that retailer 1 has the private information about retailer 2’s overconfidence in case I, thus he can
make better decisions compared with case II. Compared case I with case II, we have
. The
reason may be that retailer 2’s selling price is too high, which increases retailer 1’s product demand, thereby
improving retailer 1’s profit. Last but not least, as is shown in Fig. 5, retailer 2’s expected profit decreases as
the overconfident level increases.
It is important to note that
. It means that although the two retailers are competitors, they can
share some market information, which can make them recognize more information on market environment
and then make more rational decisions and does not necessarily result in profit loss.
In both cases I and II, retailer 2 makes a higher selling price and a lower ordering quantity because of his
overconfidence. Besides, from Fig.6, we know that
apparently, therefore, the manufacturer
can benefit from retailer 1’s knowledge of retailer 2’s overconfident level and lose considerable profit when
retailer 1 does not have the knowledge. Specifically, when retailer 1 is not aware of retailer 2’s overconfident
level, the manufacturer’s expected profit decreases sharply with the overconfident level.
Fig. 6. The effect of the overconfidence on the manufacturer’s profit
5.3 The impact of overconfidence on the supply chain
In the area of finance, overconfident decision-makers tend to overestimate their understanding of the market
environment and underestimate the uncertainty of the market, which leads to some consequences such as
unreasonable mergers and excessive investment. In a similar vein, the overconfident retailer often
overestimates the market demand, and determines a higher selling price and a lower sales volume as a
0 0.2 0.4 0.6 0.8 11750
1850
1950
2000
apro
fit
I
2
II
2
22
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
16
consequence. The overconfidence can increase retailer 2’s price, however, whether it reduces the profit of the
whole supply chain needs to be further studied.
We calculate the profit of the entire supply chain in cases I and II, and make comparisons with the
benchmark model. From Fig.7, we find that when . Meanwhile, the lower retailer 2’s
overconfident level is, the smaller the profit deviates from . Obviously, the overconfidence is not
necessarily harmful to the whole supply chain. A possible theoretical interpretation is as follows: the decisions
in the benchmark model (decentralized scenario) deviate from those in a centralized setting, thus bring about
lower profit than that in the centralized scenario, while the overconfidence may make both retailers’ pricing
and ordering decisions closer to the optimal decisions in the centralized scenario. Our conclusion is different
from the conclusion in Croson et al. (2011), they consider a two-echelon supply chain composed of one
rational manufacturer and one overconfident retailer. Besides, the demand is independent of the selling price.
They conclude that the higher the overconfident level is, the more the loss of the whole supply chain will
suffer. However, in our paper, we study a two-echelon supply chain composed of one rational manufacturer,
one rational retailer and one overconfident retailer in a duopolistic market. Besides, the demand is price-
dependent.
Fig. 7. The effect of overconfidence on profit of the whole supply chain
6. Conclusions
In this paper, we examine a duopolistic supply chain with two retailers compete in selling the same product,
one retailer is rational decision-maker, and the other retailer is overconfident. In such context, we aim to
investigate the impact of overconfidence on supply chain members’ optimal decisions and their profits under
uncertain demand, particularly in two cases: (1) the rational retailer realizes the other retailer’s
overconfidence; (2) the rational retailer is not aware of the other retailer’s overconfidence. After investigating
the effect of overconfidence and its information value by using modeling and numerical analysis, we put forth
three main findings: 1) Findings on pricing policies. In both cases, the overconfident retailer’s optimal prices
are higher than that in the benchmark model, and increase with the overconfident level. In case I, for the
rational retailer, similar to the overconfident retailer’s pricing decision, he raises his price, but the amount of
the price increasing is smaller than that of the overconfident retailer. In contrast, in case II, his pricing
0 0.2 0.4 0.6 0.8 1
5060
5070
5080
5090
a
pro
fit
I
II
N
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
17
decision is the same with that in the benchmark model since he is not aware of his rival’s overconfidence. 2)
Findings on expected profits. Compared with the benchmark model, the overconfident retailer’s expected
profit decreases in both cases, the profit loss in case II is higher than that in case I. However, the rational
retailer’s expected profits increase with the overconfident level. Moreover, for the same overconfident level,
his expected profit in case I is higher than that in case II. This is because the rational retailer has the
information about his rival’s overconfident level, which not only leads to an increase in his expected profit but
also an increase in his rival’s expected profit (compared with case II). Besides, the manufacturer’s expected
profit increases in case I while decreases in case II. This result implies that sharing overconfident information
is beneficial to each member in the supply chain. 3) Findings on supply chain performance. The retailer’s
overconfident behavior is not necessarily detrimental to the supply chain profit (compared to the benchmark
model). Previous research finds that the overconfident behavior has a negative effect on the supply chain
(Croson et al., 2011). In this paper, we complement their results by analyzing a duopolistic market, and find
that the profit of the supply chain increases in cases I and II (compared to the benchmark model).
Our analysis can be extended in several directions. First, we assume that the demand is uniformly
distributed, in reality the demand follows a variety of distributions. Thus our model can be extended to cover
other demand distribution cases, such as normal distribution. Besides, our paper studies the impact of
overconfidence in the supply chain members’ optimal decisions and profits. In fact, in a duopolistic supply
chain with overconfidence, how the manufacturer designs incentive to motivate the retailers and coordinate
the supply chain is also an interesting topic.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
18
Acknowledgements
The authors sincerely thank the editor and the anonymous reviewers for their constructive and important
comments on the paper. Lei Xu's research is partially supported by National Natural Science Foundation of
China, Grant no. 71672125, 71302005. Peng Du's research is partially supported by National Natural Science
Foundation of China, Grant no. 71702130, and major research project of National Natural Science Foundation
of China, Grant No. 91646117. Xiaoran Shi's research is partially supported by National Natural Science
Foundation of China, Grant no. 71701151.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
19
Appendix
Proof of Theorem 1. Let’s solve retailer 2’s optimal pricing first. Define
( ) , and then retailer 2’s anticipated profit can be rephrased as
(
) ∫ (
( ) (
)) ( )
∫ (
( )
)
( )
( ) .
Solving the first-order condition of retailer 2’s profit function with respect to and
, we can obtain
that
( )( ) (
)
( ) , (11)
( ) ( )
. (12)
With the same logic above, we analyze retailer 1’s optimal pricing in the view of retailer 2. Solving the
first-order condition of retailer 1’s profit function with respect to and
, respectively, we can obtain
( )( ) (
)
( ), (13)
( ) ( )
. (14)
Jointly solving the equations (10)-(13), optimal values of ,
, , and
could be derived. Since
the complexity of the optimal solutions, we only reserve the succinct form shown in Theorem 1. Q.E.D
Proof of Theorem 2. Similar to Theorem 1, we define
(
). Therefore, retailer
1’s expected profit can be rewritten as
(
) ∫ (
(
)) ( )
∫ (
) ( )
.
Solving the first-order condition of retailer 1’s profit function with respect to and
, respectively, we
can obtain
( )( ) (
)
( ) ,
(
) ( )
,
where is retailer 2’s optimal price predicted by retailer 1. Q.E.D
Proof of Theorem 3. Proofing of the Theorem 3 is similar to the Theorem 1 and 2 in case I. We define
(
), then retailer 1’s anticipated profit function is
(
) ∫ (
(
)) ( )
∫ (
) ( )
.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
20
Solving the first-order condition of retailer 1’s profit function with respect to and
, respectively,
we can obtain
( )( ) (
)
( ),
(
) ( )
.
At this moment, retailer 2’s anticipated profit function in the view of retailer 1 is
(
) ∫ (
(
)) ( )
∫ (
) ( )
.
Similarly, by taking the first-order condition of retailer 2’s expected function with respect to and ,
respectively,
( )( ) (
)
( ) ,
(
) ( )
. Q.E.D
Proof of Corollary 1. According to the result obtained from Theorem 1, ( ) ( )
, that is
. Notice that
( )( ) (
)
( ), thus we have
( )( )( )
( )
. Therefore, decreases with respect to , i.e., the higher the
overconfident level is, the higher price retailer 2 charges, and vice versa. Q.E.D
Proof of Corollary 2. In case I,
( )( )
( )
. After simplification, we obtain
(
) ( )( )
( )
. Let
and , it can be
rewritten as
( )( )
.
Taking the first-order and second-order derivative of with respect to , respectively, we can obtain
( )( )
,
( )( )
.
Since
, then is a strictly convex function with respect to . Thus we can obtain that is
strictly convex with respect to . Q.E.D
Proof of Corollary 3. Since retailer 2’s decisions in case I are same with that in case II, we only analyze the
relationship between ( ) and (
). In the view of retailer 2, his anticipated profit is
(
) ∫ (
( ) (
)) ( )
∫ (
( )
)
( )
( )
( )(
( ) )
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
21
∫
∫ (
)
∫
( )(
( ) )
(
)(( )
) (
)
( )
Similarly, retailer 2’s actual expected profit is
(
) ( )(
)
(
)(( )
) (
)
( ) .
Therefore, the difference of expected profit is
(
) (
)
( ) (
) (
)( ) (
)(
) ( )(
)
( )
( )(
)
(
)(
)
( )
[ ( )((
))
( )(
)
( )]
( )*(
)(
) ( )(
)
( )(( )(
) ( )
( )( ))+
( )[(
)(
) ( )(
)
( )
(( )( ) (
)
( )( ))]
( ) (
)(
) ( )( )( ) .
Solving the first-order derivative of the difference with respect to , we can obtain
( )( )
( )
(
) ( ) ( )
( )( )( )
( )
( )(
)
( )
(
) ( ) ( )
( )( )( )
( )
( )
( )( )
( )
( )(
)
( )
(
) ( ) ( )
( )( )( )
( )
( )
( )( )
( )( )(
)
( )
(
) ( ) ( )
( )( )( )
( )
Therefore, the difference of expected profit is a decreasing function with respect to . That is, the profit
loss increases with the overconfident level. Q.E.D
Proof of Corollary 4. Since retailer 1 realizes retailer 2’s overconfident level in case I, it can be obviously
derived that ( ) (
).
We simplify retailer 1’s profit in case II, as shown in the following equation.
a
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
22
( ) (
)(
) (
)(
) ( )
( )
( )
Similarly, in the benchmark model, retailer 1’s expected profit is
( ) (
)(
) (
)(( )
) (
) ( )
(
)
( )
where
( )( ) (
)
( ) and
(
) ( )
, .
Note that
and
, we can easily obtain ( ) (
). Therefore, we have ( )
( ) (
). Q.E.D.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
23
References
Agatz, Niels A.H., Fleischmann, M., van Nunen, Jo A.E.E. (2008). E-fulfillment and multi-channel distribution – A
review. European Journal of Operational Research, 187(2), 339–356.
Bernstein, F., & Federgruen, A. (2003). Pricing and replenishment strategies in a distribution system with
competing retailers. Operations Research, 51(3): 409-426.
Camerer, C., & Lovallo, D. (1999). Overconfidence and excess entry: An experimental approach. The American
Economic Review, 89(1), 306-318.
Chandrashekaran, R., & Grewal, D. (2006). Anchoring effects of advertised reference price and sale price-the
moderating role of saving presentation format. Journal of Business Research, 59(10), 1063-1071.
Chen, Y., Fang, S., & Wen, U. (2013). Pricing policies for substitutable products in a supply chain with internet
and traditional channels. European Journal of Operational Research, 224, 542-551.
Choi, S. C. (1991). Price competition in a channel structure with a common retailer. Marketing Science, 10(4): 271-
296.
Croson, D., Croson, R., & Ren, Y. (2011). How to manage an overconfident newsvendor. Working paper,
University of Texas at Dallas, Richardson.
Cui, T., Raju, J., & Zhang, Z. (2007). Fairness and channel coordination. Management science, 53(8), 1303-1314.
Du, P., Xu, L., Chen, Q. & Tsai S. B. (2016). Pricing Competition on Innovative Product between Innovator and
Entrant Imitator Facing Strategic Customers. International Journal of Production Research, 1-19. (doi:
http://dx.doi.org/10.1080/00207543.2015.1134837)
Fowler, James H.; Johnson, Dominic D. P. (2011-01-07). "On Overconfidence". Seed Magazine. ISSN 1499-0679.
Gervais, S., Heaton, J. B., & Odean, T. (2002). The positive role of overconfidence and optimism in investment
policy. Working Paper.
Hawkins, T. G., Wittmann, C. M., & Beyerlein, M. M. (2008). Antecedents and consequences of opportunism in
buyer- supplier relations: Research synthesis and new frontiers. Industrial Marketing Management, 37(8), 895-909.
Heaton, J. B. (2002). Managerial optimism and corporate finance. Financial Management, 31(2), 33-45.
Ho,T., Su, X., & Wu, Y. (2014). Distributional and peer-induced fairness in supply chain contract design.
Production Operations Management, 23(2), 161-175.
Ingene, C. A., & Parry, M. E. (1995). Channel coordination when retailers compete. Marketing Science, 14(4), 360-
377.
Klepper, S., & Sleeper, S. (2005). Entry by spinoffs. Management Science, 51(8), 1291-1306.
Lau, A. H. L., &Lau, H. S. (2003). Effects of a demand-curve’s shape on the optimal solutions of a multi-echelon
inventory/pricing model. European Journal of Operational Research, 147(3), 530-548.
Li M., Petruzzi N.C. & Zhang J. (2016). Overconfident competing newsvendors. Management Science, published
online (doi: http://dx.doi.org/10.1287/mnsc.2016.2469)
Liu, W., Song, S., & Wu, C. (2013).Impact of loss aversion on the newsvendor game with product substitution.
International Journal of Production Economics, 141(1), 352-359.
Malmendier, U., & Tate G. (2005). CEO overconfidence and corporate investment. Journal of Finance, 60(6),
2661-2700.
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIP
T
24
Marn, M., & Rosielb, R. (1992). Managing price, gaining profit. Harvard business review, 84-94.
Mills, E. S. Uncertainty and price theory. (1959). The Quarterly Journal of Economics, 73(1), 116-130.
Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological review, 115(2), 502-517.
Mukhopadhyay, S. K., Yue, X., & Zhu, X. (2011).A Stackelberg model of pricing of complementary goods under
information asymmetry. International Journal of Production Economics, 134(2), 424-433.
Parthasarathi, G., Sarmah, S. P., & Jenamani, M. (2011). Supply chain coordination under retail competition using
stock dependent price-setting newsvendor framework. Operational Research, 11(3), 259-279.
Petruzzi, N. C., & Dada, M. (1999). Pricing and the newsvendor problem: A review with extensions. Operations
Research, 47(2), 183-194.
Plous, S. (1993). The psychology of judgment and decision making. New York: McGraw-Hill.
Proeger, T., & Meub, L. (2014). Overconfidence as a social bias: Experimental evidence. Economics Letters,
122(2), 203-207.
Ren Y., Croson D.C. & Croson R. (2017). The overconfident newsvendor. Journal of the Operational Research
Society, 68: 496-506.
Ren, Y., & Croson, R. (2013). Overconfidence in newsvendor orders: An experimental study. Management
Science, 59(11), 2502-2517.
Wang, Z., Zhang, Z., Li, C., Xu, L., & You, C. (2015). Optimal ordering and disposing policies in the presence of
an overconfident retailer: A Stackelberg game. Mathematical Problems in Engineering.
Weinstein N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social
Psychology, 39(5), 806-820.
Xiao, T., & Qi, X. (2008). Price competition, cost and demand disruptions and coordination of a supply chain with
one manufacturer and two competing retailers. Omega, 36(5), 741-753.
Yang, D., Xiao, T., & Shen, H. (2009). Pricing, service level and lot size decisions of a supply chain with risk-
averse retailers: implications to practitioners. Production Planning and Control, 20(4), 320-331.
Yu Z., Yan H, Edwin Cheng T C. (2001). Benefits of information sharing with supply chain partnerships. Industrial
Management & Data Systems, 101(3), 114-121.